| Age | Commit message (Collapse) | Author |
|---|
|
We previously crated a HandleScope in the JNI transition
frame to hold references passed as jobject (jclass, etc.)
to the native function and these references were actually
spilled twice during the transition.
We now construct the jobject as a pointer to the reference
spilled in the reserved out vreg area in the caller's frame.
And the jclass for static methods is just a pointer to the
method's declaring class. This reduces the amount of work
required in the JNI transition, both on entry (in compiled
stubs) and exit (in JniMethodEnd*).
Some additional work is required when GC visits references
of a native method as we need to walk over the method's
shorty which was unnecessary for a HandleScope.
Also fix Thread::InitStackHwm() to calculate correct stack
size needed by the new Thread::IsJniTransitionReference().
The results for StringToBytesBenchmark on blueline little
cores running at fixed frequency 1420800 are approximately
arm64 (medians from 3 runs) before after
timeGetBytesAscii EMPTY 447.33 436.86
timeGetBytesIso88591 EMPTY 440.52 431.13
timeGetBytesUtf8 EMPTY 432.31 409.82
arm (medians from 3 runs) before after
timeGetBytesAscii EMPTY 500.53 490.87
timeGetBytesIso88591 EMPTY 496.45 495.30
timeGetBytesUtf8 EMPTY 488.84 472.68
Test: m test-art-host-gtest
Test: testrunner.py --host
Test: testrunner.py --host --gcstress
Test: testrunner.py --host --jit-on-first-use
Test: testrunner.py --host --jit-on-first-use --gcstress
Test: run-gtests.sh
Test: testrunner.py --target --optimizing
Test: boots.
Bug: 172332525
Change-Id: I658f9d87071587b3e89f31c65feca976a11e9cc2
|
|
And spill stack arguments directly to the right location.
Do not spill to the reserved space in the caller's frame.
Preliminary Golem results for art-opt-cc:
x86 x86-64 arm arm64
NativeDowncallCritical6: n/a +14.3% +17.2% +26.1%
(x86 seems to be currently providing results that are worse
than interpreter, so something is not working.)
Test: Additional tests in 178-app-image-native-method test.
Test: m test-art-host-gtest
Test: testrunner.py --host --optimizing
Test: aosp_taimen-userdebug boots.
Test: run-gtests.sh
Test: testrunner.py --target --optimizing
Bug: 112189621
Change-Id: I709c52ab2585a8f5f441f53ad2bf4a01d2b25dca
|
|
The calling convention no longer describes entry spills as
spilling is the JNI compiler's responsibility. This allows
future improvements, such as spilling registers directly to
the HandleScope or outgoing stack args.
Remove the notion of interprocedural scratch register from
calling conventions and let assemblers deal with all scratch
register uses. The remaining JNI assembler APIs that take
scratch registers are currently unused and can be removed.
Also fix a bug in disassembly comparison for tests; the
contents of two files were considered identical if the
second one just contained additional data.
This change fully preserves the generated code and adds TODO
comments where doing so results in weird or suboptimal code.
Test: m test-art-host-gtest
Test: testrunner.py --host --optimizing
Test: Prebuilt boot image is unchanged.
Test: aosp_taimen-userdebug boots.
Bug: 12189621
Change-Id: Ic26a670276920313cd907a6eda8d982cf0abfd81
|
|
Omit managed frame for @CriticalNative methods, do not check
for exceptions and and make a tail call when possible.
Pass the method pointer in a hidden argument to prepare for
implementing late binding for @CriticalNative methods.
This changes only the JNI compiler, Generic JNI shall be
updated in a separate change.
Performance improvements reported by Golem (art-opt-cc):
x86 x86-64 arm arm64
NativeDowncallStaticCritical6 +17% +50% +88% +139%
NativeDowncallStaticCritical +37% +32% +103% +216%
Test: m test-art-host-gtest
Test: testrunner.py --host --optimizing
Test: aosp_taimen-userdebug boots.
Test: run-gtests.sh
Test: testrunner.py --target --optimizing
Bug: 112189621
Change-Id: I5758c8f478627f2eee8f615b4537a907c211b9f8
|
|
Remove 'virtual' specifier on methods already bearing the 'override'
specifier.
Test: mmma art
Change-Id: I114930969a5ca048d88de9ecd18e2c6403593e31
|
|
Remove all uses of macros 'FINAL' and 'OVERRIDE' and replace them with
'final' and 'override' specifiers. Remove all definitions of these
macros as well, which were located in these files:
- libartbase/base/macros.h
- test/913-heaps/heaps.cc
- test/ti-agent/ti_macros.h
ART is now using C++14; the 'final' and 'override' specifiers have
been introduced in C++11.
Test: mmma art
Change-Id: I256c7758155a71a2940ef2574925a44076feeebf
|
|
Replace wherever possible. ART's base/logging is now mainly VLOG
and initialization code that is unnecessary to pull in and makes
changes to verbose logging more painful than they have to be.
Test: m test-art-host
Change-Id: I3e3a4672ba5b621e57590a526c7d1c8b749e4f6e
|
|
Memory needed to compile the two most expensive methods for
aosp_angler-userdebug boot image:
BatteryStats.dumpCheckinLocked() : 25.1MiB -> 21.1MiB
BatteryStats.dumpLocked(): 49.6MiB -> 42.0MiB
This is because all the memory previously used by Scheduler
is reused by the register allocator; the register allocator
has a higher peak usage of the ArenaStack.
And continue the "arena"->"allocator" renaming.
Test: m test-art-host-gtest
Test: testrunner.py --host
Bug: 64312607
Change-Id: Idfd79a9901552b5147ec0bf591cb38120de86b01
|
|
CriticalNative methods shall not be suspended and hence do not
require MR to be refreshed in compiled JNI code.
This change is for ARM and ARM64 only.
Impact on Critical Native benchmarks times (median of 10 runs,
lower is better):
* angler-userdebug - ARMv7
** All cores
NativeDowncallStaticCritical -2.78%
NativeDowncallStaticCritical6 -1.79%
** Little cores only
NativeDowncallStaticCritical -1.66%
NativeDowncallStaticCritical6 -1.27%
** Big cores only
NativeDowncallStaticCritical -2.66%
NativeDowncallStaticCritical6 -1.70%
* angler-userdebug - ARMv8
** All cores
NativeDowncallStaticCritical -3.52%
NativeDowncallStaticCritical6 -1.79%
** Little cores only
NativeDowncallStaticCritical -1.63%
NativeDowncallStaticCritical6 -1.27%
** Big cores only
NativeDowncallStaticCritical -3.87%
NativeDowncallStaticCritical6 -1.75%
Test: m test-art-target
Test: m test-art-target with tree built with ART_USE_READ_BARRIER=false
Test: m test-art-host-gtest
Test: ARM64 device boot test
Test: ARM device boot test
Bug: b/37707231
Change-Id: I95d61b9ecde0afffdd5fd44763b19caa06025ec8
|
|
Generate run-time code in the Optimizing compiler checking that
the Marking Register's value matches `self.tls32_.is.gc_marking`
in debug mode (on target; and on host with JIT, or with AOT when
compiling the core image). If a check fails, abort.
Test: m test-art-target
Test: m test-art-target with tree built with ART_USE_READ_BARRIER=false
Test: ARM64 device boot test with libartd.
Bug: 37707231
Change-Id: Ie9b322b22b3d26654a06821e1db71dbda3c43061
|
|
Static method dispatch via JNI requires a read barrier
for the ArtMethod::GetDeclaringClass() load before adding it to the
JNI StackHandleScope.
We used to call ReadBarrierJni unconditionally but add a branch
to skip calling it if the GC is not currently in the marking phase.
Test: ART_USE_READ_BARRIER=true make test-art-host test-art-target
Bug: 30437917
Change-Id: I4f505ebde17c0a67209c7bb51b3f39e37a06373a
|
|
Will be used in upcoming CLs regarding VDEX and VerifierDeps.
Test: m test-art-host
Change-Id: I68e611a4a52246c2bdf45eab7c61f3212908afd4
|
|
Extract macro assembler functionality used by the JNI compiler from
the assembler interface. Templatize the new interface so that
type safety ensures correct usage.
Change-Id: Idb9f56e5b87e43ee6a7378853d8a9f01abe156b2
Test: m test-art-host
|
